INDEX

Explanations

code-related terms and operations involving variable references and data structures

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 34-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_34/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.34.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

\{\\

-0.43

 ModelExpression

-0.39

 berätt

-0.33

 banget

-0.31

 beginnetje

-0.30

 juuri

-0.29

 justiça

-0.28

BASEPATH

-0.28

 śmier

-0.27

 smrti

-0.27

POSITIVE LOGITS

 שוליים

0.72

BagConstraints

0.66

CultureInfo

0.65

 referrerpolicy

0.61

HtmlAttribute

0.59

 extras

0.58

 Chwiliwch

0.57

StandardCharsets

0.56

waitKey

0.55

sika

0.54

Activations Density 0.506%

code-related terms and operations involving variable references and data structures

No Comments

No Known Activations